Writing programming and writing machine learning models
Based on the different machine learning models, a large number of characteristic variables are used to predict the fluctuation of the underlying asset price, and the prediction results are evaluated. Machine learning models include, but are not limited to, Xgboost, GBDT, lstm and other classical learning models. Assets to be studied include: stocks, bonds, commodities and other configurable assets. The characteristic variables include macroeconomic variables, industry variables, and price sequences. For threaded steels in commodities, we will provide supply, demand, cost, and so on, a specific characteristic variable of the breed. Contestants need to take into account the characteristics of the time series in the data, as well as the organization of data between different frequencies, and through the effective feature extraction method, to build the underlying price fluctuation prediction model. We will give specific asset indices and possible characteristic variables, and contestants need to explore the predictive models of these assets.
Evaluation index:
We use the normalized RMS error (uniformed root mean squared error, URMSE) to measure the predictive error for a single sequence:
The contestants predict the sequence of Y1, Y2, Y3 and Steel, and calculate the URMSE1, URMSE2, URMSE3 and URMSE4 respectively, and the final total evaluation index = (URMSE1+URMSE2+URMSE3+URMSE4)/ 4, according to the overall evaluation indicators from low to high ranking.
Data description
1, x sequence _train.xls, x sequence _test.xls, iron and steel x_train.xls, steel data _test.xls provided the training set and test set of macro-economic variables, industry variables and other characteristics of variables, players to build effective features.
2, Y_train.xls contains 4 sheet, respectively, the training concentrated Y1, Y2, Y3, steel in different time periods of price fluctuations.
3. Sample submission:
3.1. Submit a TXT file with UTF-8 without BOM encoding, and submit a TXT file altogether.
3.2.y1,y2,y3, the price forecast for steel is divided into four modules, each of which starts with a single line of Y1,y2,y3,iron string identification.
3.3. Predictions must be made at a given point in time, with a \ t partition between date and price, and no missing or extra data.
Format:
Y1
Date1 Price
Date2 Price
...
Y2
Date1 Price
...
Y3
...
Iron
...
3.4 The specific format can refer to Submit_sample.txt (note: Submit_sample.txt only gives the date, please add \ t and the corresponding predicted value after the date of each line when submitting)
Our Direction field: Window Programming numerical algorithm AI Artificial Intelligence financial statistical Metrology analysis Big Data network programming Web programming Communication Programming game Programming Multimedia Linux plug-in programming API image processing embedded/Microcontroller database programming console process and thread Network security assembly language Hardware programming software Design Engineering Standard Rules. The generation of programming languages or tools including, but not limited to, the following ranges:
C/c++/c# Write
Java Write generation
It generation
Python writes
Tutoring Programming Jobs
The MATLAB Generation writes
Haskell writes
Processing Write
Linux Environment Setup
Rust Generation Write
Data Structure assginment Data structure generation
MIPS Generation Writing
Machine Learning Job Writing
Oracle/sql/postgresql/pig database Generation/Generation/Coaching
Web development, Web development, Web site jobs
Asp. NET Web site development
Finance insurace Statistics Statistics, regression, iteration
Prolog write
Computer Computational Method Generation
Because of professional, so trustworthy. If necessary, please add qq:99515681 or e-mail:[email protected]
: Codinghelp
Write programming, write machine learning models, write AI Python on behalf of